FormalPara Take-home message

After 1 year of follow-up, patients with more severe COVID-19, defined as need for mechanical ventilation during hospitalisation, had significantly lower health-related quality of life and worse results for mortality, major cardiovascular events, re-hospitalisation, new disabilities in instrumental activities of daily living, anxiety and post-traumatic stress symptoms, and return to work or study than COVID-19 patients who did not need mechanical ventilation during hospitalisation

Introduction

Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2 infection, has affected millions of people around the world. Brazil has been severely hit by the pandemic with the number of cases surpassing 35 million, with more than 680,000 deaths from COVID-19 by November 2022 [1]. Although a large amount of comprehensive data on acute symptoms and clinical management have been published, the long-term effects of COVID-19 remain unclear [2]. Recent studies have drawn attention to an increasing number of people experiencing prolonged symptoms following the acute phase of COVID-19 [3,4,5,6,7]. However, our knowledge of the long-term impact of acute COVID-19 severity on relevant outcomes, such as quality of life, cardiovascular events, new functional disabilities, and mental health symptoms, is rather limited. Notably, this evidence gap may constitute a barrier to understanding epidemiology, risk factors, and the natural history of post-COVID-19 disabilities, precluding the implementation of effective prevention and rehabilitation strategies. Accordingly, we conducted the Coalition VII prospective cohort study to investigate whether acute COVID-19 severity is associated with 1-year quality of life.

Methods

Study design and follow-up

The rationale and design of the Coalition VII (NCT04376658) have been published previously [8]. Briefly, this is a multicentre prospective cohort study nested in five randomised clinical trials originally designed to assess the effects of specific COVID-19 treatments in hospitalised adult patients in Brazil [9,10,11,12,13]. Survivors were followed up for 1 year by means of structured and centralised telephone interviews conducted at 3, 6, 9, and 12 months after enrolment in clinical trials that compose this study. For patients with communication difficulties, the follow-up interviews were conducted with their proxy.

All randomised clinical trials that compose the present cohort study, including their amendments for 1-year telephone follow-up, were approved by Brazil’s National Ethics Committee (Electronic Supplemental Material, ESM 1). Written informed consent was obtained from participants or their proxies at the time of enrolment during hospital stay. Participants were re-consented during the first telephone call.

Participants

This study included patients aged ≥ 18 years requiring hospitalisation for proven or suspected SARS-CoV-2 infection and meeting eligibility criteria for Coalition I (hospitalised patients with suspected or confirmed COVID-19 who were receiving either no supplemental oxygen or a maximum of 4 L/min of supplemental oxygen) [9], Coalition II (hospitalised patients with suspected or confirmed COVID-19 and at least one additional severity criteria: use of oxygen supplementation > 4 L/min; use of high-flow nasal cannula; use of non-invasive ventilation; or use of mechanical ventilation) [10], Coalition III (hospitalised patients with suspected or confirmed COVID-19 with moderate-to-severe acute respiratory distress syndrome, ARDS) [11], Coalition IV (hospitalised patients with confirmed COVID-19 and elevated serum d-dimer concentration) [12], and Coalition VI (hospitalised patients with confirmed COVID-19 who were receiving supplemental oxygen or mechanical ventilation and had abnormal levels of at least two serum biomarkers: C-reactive protein, d-dimer, lactate dehydrogenase, or ferritin) [13] randomised clinical trials. Complete information on the objectives, eligibility criteria, and period of enrolment for each trial that composes the present cohort is provided in the ESM 2. We excluded patients who died during hospitalisation, who lacked a telephone contact, or who refused or withdrew consent to participate.

Patients with a positive reverse transcription-polymerase chain reaction (RT-PCR) test for SARS-CoV-2 were considered proven cases. Suspected cases were defined according to the Brazilian Ministry of Health criteria: presence of fever and at least one respiratory sign or symptom (e.g. cough, shortness of breath, nasal congestion, difficulty swallowing, sore throat, oxygen saturation < 95%, signs of cyanosis, intercostal retraction, and dyspnoea) and patients from an endemic region, or travelling from an endemic region in the last 14 days, or in contact with a suspected or confirmed case in the last 14 days [14].

Acute disease severity

Acute disease severity was determined by the highest score on a modified six-point ordinal scale [15] during hospital stay, which consisted of the following categories: a score of 1 indicated not hospitalised; 2, hospitalised but no supplemental oxygen needed; 3, hospitalised and receiving supplemental oxygen; 4, hospitalised and receiving high-flow nasal cannula oxygen therapy or non-invasive ventilation; 5, hospitalised and receiving mechanical ventilation; and 6, death. Patients classified as score 1 or 6 were not enrolled in this study.

Outcomes

Primary outcome

The primary outcome was the health-related quality of life utility score measured at 1 year after enrolment with the EuroQol five-dimension three-level (EQ-5D-3L) questionnaire [16]. The EQ-5D-3L consists of a descriptive system with five dimensions of health-related quality of life (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression), and each dimension has three levels (no problems, some problems, and extreme problems). The utility score derived from the descriptive system for the Brazilian population ranges from − 0.17 (where 0 is a health state equivalent to death; negative values are valued as worse than death) to 1 (best health state) [16]. The estimated minimal clinically important difference of EQ-5D-3L is 0.03 [17], and the mean value for the Brazilian population is 0.82 [18]. Patients who died during follow-up received a score of 0 on all follow-ups after the event.

Secondary outcomes

Secondary outcomes included EQ-5D-3L utility scores measured at 3, 6, and 9 months after enrolment, and all-cause mortality, major cardiovascular events (non-fatal stroke, non-fatal myocardial infarction, and cardiovascular death), re-hospitalisations, new disabilities in instrumental activities of daily living assessed by the Lawton and Brody instrumental activities of daily living scale [19] (any impairment, moving from independent to partially dependent or from partially dependent to totally dependent, in at least one of the following domains: telephone use, transportation, shopping, responsibility for own medications, and ability to handle finances) relative to 1 month before hospitalisation, dyspnoea assessed by the modified Medical Research Council dyspnoea scale [20], need for home ventilatory support (oxygen, non-invasive ventilation, or mechanical ventilation), anxiety and depression symptoms assessed by the Hospital Anxiety and Depression Scale (scores > 7 indicate possible cases of anxiety or depression) [21], post-traumatic stress symptoms assessed by the Impact of Event Scale-Revised (scores > 33 indicate possible cases of post-traumatic stress disorder) [22], and return to work or study at 3, 6, 9, and 12 months after enrolment.

Statistical analysis

The baseline characteristics and outcomes of participants were presented as median (interquartile range, IQR) or mean (standard deviation, SD) for continuous variables, and as counts and percentages for categorical variables. Generalised estimating equations adjusted for age, sex, number of comorbidities, and the trial in which the patient was enrolled were used to estimate adjusted differences and 95% confidence intervals (CIs) for the association between acute disease severity and outcomes. Variables used for covariate adjustment were selected based on their association with health-related quality of life [18, 23] and to account for the cluster effect.

Sensitivity analyses for the primary outcome considering only patients with proven COVID-19, adjusting analyses by pre-morbid functional dependency, and using mixed effects model for statistical modelling were performed to assess the consistency of the findings. Additionally, we assessed the association between acute disease severity and all-cause mortality using a frailty model (adjusted for age, sex, number of comorbidities, and the trial in which the patient was enrolled) to account to censored data. Additional post hoc analyses are described in the ESM 3. We used R version 3.6.2 (R Foundation for Statistical Computing) for all statistical analyses. All tests were two-tailed, and p values < 0.05 were considered statistically significant. We did not adjust the confidence interval widths and p values of secondary outcomes for multiple testing.

Results

A total of 1508 patients (including 1332 patients with confirmed COVID-19) at 84 sites were enrolled between March 29, 2020, and February 26, 2021 (Fig. 1). Amongst the 1508 enrolled patients, 36 (2.4%) died before completing the 1-year follow-up. Thus, 1472 patients were available for the 1-year follow-up, which was completed on March 11, 2022. Amongst the 1472 patients eligible for the 1-year follow-up, 1257 (85.4%) were assessed. Data on the primary outcome were available for 1156 participants (1120 survivors and 36 dead patients).

Fig. 1
figure 1

Flow diagram of post-hospitalisation survivors of COVID-19. RCT randomised clinical trial. 11120 survivors and 36 dead patients

Full size image

Characteristics of participants

Table 1 shows the characteristics of participants. The median age of participants was 53.2 years (IQR 42.3–64), and 917 (60.8%) were men. The most common comorbidities were hypertension (681 patients [45.2%]), obesity (451 patients [30.2%]), and diabetes (365 patients [24.2%]). Pre-morbid functional dependency was present in 615/1315 (46.8%). According to the highest score on the ordinal severity scale during hospital stay, 688 (45.6%) were categorised as score 2, 394 (26.1%) as score 3, 94 (6.2%) as score 4, and 332 (22%) as score 5. The median duration of mechanical ventilation was 10 days (IQR 6–18). The median length of hospital stay was 8 days (IQR 5–19.8). The baseline characteristics of patients assessed for the primary outcome and patients with missing values for the primary outcome were similar (ESM Table S1).

Table 1 Characteristics of enrolled patients
Full size table

Quality of life

The results of quality of life are showed in Table 2. All 1156 participants with available data for the EQ5D-3L at 1 year were included in the primary outcome analysis. At 1 year, the mean EQ-5D-3L utility score for the entire cohort was 0.8 (SD, 0.24). Patients with severity score 5 had lower mean EQ-5D-3L utility scores than those with score 2 (0.7 vs 0.84; adjusted difference, − 0.1 [95% CI − 0.15 to − 0.06]). The mean EQ-5D-3L utility scores of patients with severity scores 3 and 4 did not differ significantly from those of patients with score 2. The mean EQ-5D-3L utility scores at 3, 6, and 9 months were also lower for severity score 5 vs score 2 patients (Fig. 2A). Compared to patients who did not need mechanical ventilation (severity scores 2–4), patients who needed mechanical ventilation (severity score 5) had lower mean EQ-5D-3L utility scores at 1 year (0.7 vs. 0.8; adjusted difference, − 0.07 [95% CI − 0.11 to − 0.04]; Fig. 2B).

Table 2 Health-related quality of life, mortality, cardiovascular events and re-hospitalisation among post-hospitalisation COVID-19 patients
Full size table
Fig. 2
figure 2

Effect of COVID-19 severity on EQ-5D-3L utility scores at 3, 6, 9, and 12 months

Full size image

Mortality, cardiovascular events and re-hospitalisations

Patients with severity score 5 had higher 1-year incidence of all-cause mortality (7.9% vs 1.2%; adjusted difference, 7.1% [95% CI 2.5–11.8%]), major cardiovascular events (5.6% vs 2.3%; adjusted difference, 2.6% [95% CI 0.6–4.6%]), and re-hospitalisations (24.4% vs 19.6%; adjusted difference, 4.2% [95% CI 1–7.4%]) than score 2 patients (Table 2).

Regarding the 1-year incidence of components of major cardiovascular events (ESM Table S2), compared with severity score 2 patients, score 5 patients had higher incidence of non-fatal myocardial infarction (adjusted difference, 2% [95% CI 1–3%]) and cardiovascular death (adjusted difference, 1% [95% CI 0.2–2%]). The incidence of non-fatal stroke did not differ significantly between severity score groups.

New disabilities, home ventilatory support, mental health symptoms and return to work or study

Severity score 5 patients had higher 1-year incidence of new disabilities in instrumental activities of daily living (40.4% vs 23.5%; adjusted difference, 15.5% [95% CI 8.5–22.5%]), higher 1-year prevalence of home ventilatory support (3.4% vs 1.3%; adjusted difference, 2.1% [95% CI 0.6–3.6%]), anxiety symptoms (24.7% vs 17.5%; adjusted difference 6.5% [95% CI 3.1–9.8%]) and post-traumatic stress symptoms (14% vs 7.1%; adjusted difference, 6.4%, [95% CI 4.6–8.2%]), and lower 1-year incidence of return to work (88.1% vs 97.5%; adjusted difference, − 7.4% [95% CI − 11.8 to − 2.9%]) or study (88.9% vs 96.9%; adjusted difference, − 10.5% [95% CI − 16.2 to − 4.9%]) than score 2 patients (Table 3).

Table 3 New disabilities, home ventilatory support, mental health symptoms and return to work or study among post-hospitalisation COVID-19 patients
Full size table

Regarding the 1-year prevalence of components of home ventilatory support (ESM Table S3), oxygen use was higher for severity score 5 vs score 2 patients (4.8% vs 0.3%; adjusted difference, 4.5% [95% CI 3.6–5.5%]); the prevalence of non-invasive ventilation and mechanical ventilation did not differ significantly between severity score groups.

Additional analyses

The results of the sensitivity analyses for primary outcome were consistent with the results of main analysis (ESM Tables S4–S6). The effects of acute disease severity on the hazard of 1-year all-cause mortality, as assessed by a frailty model, were also consistent with the findings of the main analysis (ESM Table S7).

The 1-year prevalence of dyspnoea was higher for severity score 5 vs score 2 patients (ESM Table S8). Dyspnoea was more severe in patients with severity score 5 than in those with score 2 at 3, 6, 9 and 12 months (ESM Fig S1). Patients with severity score 5 scored worse in the EQ-5D-3L domains of mobility, usual activities, pain/discomfort, and anxiety/depression than did those with severity scores 2 at 1 year (ESM Fig. 2). No association was observed between duration of mechanical ventilation and EQ-5D-3L utility scores at 1 year (ESM Fig. 3 and Table S9). Major cardiovascular events, re-hospitalisations, new disabilities in instrumental activities of daily living, dyspnoea, home ventilatory support, anxiety, depression and post-traumatic stress symptoms were associated with lower EQ-5D-3L utility scores at 1 year (ESM Table S10).

The results of comparison between patients who did not need mechanical ventilation and those who needed mechanical ventilation on secondary outcomes were consistent with main analyses (ESM Table S11).

Discussion

In this cohort study, we observed that, after 1 year of follow-up, patients with more severe COVID-19, defined as need for mechanical ventilation during hospitalisation, had lower health-related quality-of-life utility scores and worse results for mortality, major cardiovascular events, re-hospitalisation, new disabilities in instrumental activities of daily living, dyspnoea, anxiety and post-traumatic stress symptoms, and returning to work or study.

We found that post-hospitalisation COVID-19 patients who had received mechanical ventilation had clinically meaningful reductions in health-related quality of life utility scores at 3, 6, 9, and 12 months compared with those not requiring mechanical ventilation. Although these scores improved during the 1-year follow-up period, they were still below the mean value for the Brazilian population at 1-year. This is consistent with data from previous ARDS and long-term intensive care unit (ICU) follow-up studies. For example, Herridge et al. [24] demonstrated that, although survivors of ARDS improved their quality-of-life scores during the long-term follow-up, the mean score on the physical component of the 36-item Short-Form Health Survey at 5 years remained approximately 1 SD below the mean score for an age-matched and sex-matched control population. Similarly, Hofhuis et al. [25] showed that the health-related quality of life of medical–surgical ICU survivors remained impaired compared with their pre-admission values and with an age-matched reference population after 5 years, suggesting that the post-ICU COVID-19 and non-COVID-19 patients might have the same rehabilitation needs at the long-term. Notably, oxygen delivered by mask or nasal prongs or by non-invasive ventilation and use of high-flow nasal cannula oxygen therapy were not associated with a significant reduction in health-related quality-of-life utility scores in our study, suggesting that acute strategies aimed to prevent mechanical ventilation among patients with COVID-19 might be associated with improved long-term outcomes.

Concerning physical and mental disabilities, the findings of the present study showed a higher occurrence of new functional disabilities, dyspnoea and of anxiety and post-traumatic stress symptoms in patients requiring mechanical ventilation during hospitalisation. For example, mechanical ventilation patients had twice the prevalence of post-traumatic stress symptoms than the general population in Brazil [26]. We hypothesise that these results might have contributed to the worse quality of life among patients with more severe COVID-19, which is underlined by the association between physical and mental health outcomes and reduced 1-year quality of life in our study. Accordingly, new physical and mental disabilities have been associated with reduced quality of life among survivors of critical illness [27, 28].

The association between COVID-19 severity and higher occurrence of major cardiovascular events in this study is consistent with the population-based cohort study conducted by Xie et al., [29] who showed that, beyond the first 30 days after infection, individuals with COVID-19 are at increased risk of incident cardiovascular disease, and with the literature on long-term sepsis outcomes, which shows an excess hazard of late cardiovascular events among sepsis survivors which may persist for at least 5 years following hospital discharge [30].

The higher impact of COVID-19 severity on all-cause mortality and re-hospitalisations is also a reason for concern. These data reinforce that the attributable impact of COVID-19 on hospitalisations and mortality is even higher than that associated with acute illness.

Potential explanations were considered for the association between need for mechanical ventilation and poor long-term outcomes among survivors of COVID-19. First, need for mechanical ventilation can be interpreted as a proxy for disease severity, which may lead to persistent organ dysfunction after acute SARS-CoV-2 infection, thus contributing to long-term disabilities. Accordingly, studies have shown that patients with COVID-19 requiring mechanical ventilation are more likely to have elevated inflammatory markers, more extensive lung involvement, multiple organ dysfunction, and higher in-hospital mortality [31,32,33]. Second, the supportive care required by mechanically ventilated patients with COVID-19 and mechanical ventilation-related complications might have contributed to a higher occurrence of physical and mental disabilities among survivors of COVID-19. Studies of survivors of critical illness have found an association between mechanical ventilation-related factors (such as profound sedation, neuromuscular blocking agents, corticosteroids, immobilisation, and ventilator-associated pneumonia) and worse long-term outcomes (such as ICU-acquired weakness, post-traumatic stress, post-discharge mortality, and reduced quality of life) [34,35,36,37]. Third, the unprecedented critical care capacity strain caused by the COVID-19 pandemic might have been associated with lower adherence to interventions aimed at preventing long-term disabilities among mechanically ventilated patients, such as minimising sedation and use of neuromuscular blocking agents, pain control, early mobilisation, and family presence [38].

Strengths of our study include its prospective design, large sample size, 1-year follow-up, and the assessment of patient-centred outcomes. However, this study has limitations. Although the study recruited from many hospitals, the sample was limited to one middle-income country. COVID-19 may have different effects on long-term outcomes across distinct contexts in terms of post-discharge access to rehabilitation services. We did not evaluate the pre-COVID-19 values of EQ-5D-3L, precluding the assessment of utility score variations in comparison to the pre-morbid period. We did not evaluate potentially relevant variables that could modify the association between acute COVID-19 severity and long-term outcomes, such as vaccination, infection with different SARS-CoV-2 variants, and specific treatments. The number of missing assessments for 1-year outcomes was relevant. Finally, we did not include a control group of patients without COVID-19, precluding the differentiation between specific COVID-19-mediated and critical illness-mediated effects on long-term outcomes.

Conclusions

COVID-19 patients who needed mechanical ventilation during hospitalisation have lower 1-year quality of life than COVID-19 patients who did not need mechanical ventilation during hospitalisation.